DOCSLIB.ORG

Crystalline-Comparis

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Crystalline-Comparis CRYSTALLINE COHOMOLOGY AND DE RHAM COHOMOLOGY BHARGAV BHATT AND AISE JOHAN DE JONG ABSTRACT. The goal of this short paper is to give a slightly different perspective on the comparison between crystalline co- homology and de Rham cohomology. Most notably, we reprove Berthelot’s comparison result without using pd-stratifications, linearisations, and pd-differential operators. Crystalline cohomology is a p-adic cohomology theory for varieties in characteristic p created by Berthelot [Ber74]. It was designed to fill the gap at p left by the discovery [SGA73] of `-adic cohomology for ` 6= p. The construction of crystalline cohomology relies on the crystalline site, which is a better behaved positive characteristic analogue of Grothendieck’s infinitesimal site [Gro68]. The motivation for this definition comes from Grothendieck’s theorem [Gro66] identifying infinitesimal cohomology of a complex algebraic variety with its singular cohomology (with C- coefficients); in particular, infinitesimal cohomology gives a purely algebraic definition of the “true” cohomology groups for complex algebraic varieties. The fundamental structural result of Berthelot [Ber74, Theorem V.2.3.2] is a direct p-adic analogue of this reconstruction result: the crystalline cohomology of a smooth Fp-variety X is identified with the de Rham cohomology of a lift of X to Zp, provided one exists. In particular, crystalline cohomology produces the “correct” Betti numbers, at least for liftable smooth projective varieties (and, in fact, even without liftability by [KM74]). We defer to [Ill94] for a detailed introduction, and connections with p-adic Hodge theory. Our goal in this note is to give a different perspective on the relationship between de Rham and crystalline coho- mology. In particular, we give a short proof of the aforementioned comparison result [Ber74, Theorem V.2.3.2]; see Theorem 3.6. Our approach replaces Berthelot’s differential methods (involving stratifications and linearisations) with a resolutely Cech-theoreticˇ approach. It seems that Theorem 3.2 is new, although it may have been known to experts in the field. This theorem also appears in forthcoming work of Alexander Beilinson [Bei]. Conventions. Throughout this note, p is a fixed prime number. Our base scheme will be typically be Σ = Spec(Zp), e though occasionally we discuss the theory over Σe = Spec(Zp=p ) as well (for some e ≥ 1). All divided powers will be compatible with the divided powers on pZp. Modules of differentials on divided power algebras are compatible with the divided power structure. A general reference for divided powers and the crystalline site is [Ber74]. 1. REVIEW OF MODULES ON THE CRYSTALLINE SITE Let S be a Σ-scheme such that p is locally nilpotent on S. The (small) crystalline site of S is denoted (S=Σ)cris. Its objects are triples (U; T; δ) where U ⊂ S is an open subset, U ⊂ T is a nilpotent thickening of Σ-schemes, and δ is a divided power structure on the ideal of U in T ; the morphisms are the obvious ones, while coverings of (U; T; δ) are induced by Zariski covers of T . The structure sheaf OS=Σ of (S=Σ)cris is defined by OS=Σ((U; T; δ)) = Γ(T; OT ). e Given a Zp=p -algebra B and an ideal J ⊂ B endowed with divided powers δ, the module of differentials compatible with divided powers is the quotient of the module of Σ-linear differentials by the relations dδn(x) = δn−1(x)d(x), for 1 1 x 2 J and n ≥ 1. We simply write ΩB for this module as confusion is unlikely. The formation of ΩB commutes with 1 1 1 localisation on B, so the formula ΩS=Σ((U; T; δ)) = Γ(T; ΩT ) defines a sheaf ΩS=Σ on (S=Σ)cris. Like its classical 1 analogue, the sheaf ΩS=Σ can also be described via the diagonal as follows. Given an object (U; T; δ) of (S=Σ)cris, let (U; T (1); δ(1)) be the product of (U; T; δ) with itself in (S=Σ)cris: the scheme T (1) is simply the divided power envelope of U ⊂ T ×Σ T , with δ(1) being the induced divided power structure. The diagonal map ∆ : T ! T (1) is a closed immersion corresponding to an ideal sheaf I with divided powers, and we have 1 [2] ΩS=Σ((U; T; δ)) = Γ(T; I=I ); [2] i 1 where I denotes the second divided power of I. For i ≥ 0 we define ΩS=Σ as the i-th exterior power of ΩS=Σ. An OS=Σ-module F on (S=Σ)cris is called quasi-coherent if for every object (U; T; δ), the restriction FT of F to the i Zariski site of T is a quasi-coherent OT -module. Examples include OS=Σ and ΩS=Σ for all i > 0. 1 An OS=Σ-module F on (S=Σ)cris is called a crystal in quasi-coherent modules if it is quasi-coherent and for every morphism f :(U; T; δ) ! (U 0;T 0; δ0) the comparison map ∗ cf : f FT 0 !FT i is an isomorphism. For example, the sheaf OS=Σ is a crystal (by fiat), but the sheaves ΩS=Σ, i > 0 are not crystals. Given a crystal F in quasi-coherent modules and an object (U; T; δ), the projections define canonical isomorphisms ∗ c1 c2 ∗ pr1FT −!FT (1) − pr2FT : These comparison maps are functorial in the objects of the crystalline site. Hence we obtain a canonical map 1 (1.0.1) r : F −! F ⊗OS=Σ ΩS=Σ such that for any object (U; T; δ) and any section s 2 Γ(T; FT ) we have [2] c1(s ⊗ 1) − c2(1 ⊗ s) = r(s) 2 I=I ⊗OS=Σ FT (1): Transitivity of the comparison maps implies this connection is integrable, hence defines a de Rham complex 1 2 F ! F ⊗OS=Σ ΩS=Σ ! F ⊗OS=Σ ΩS=Σ !··· : We remark that this complex does not terminate in general. 2. THE DE RHAM-CRYSTALLINE COMPARISON FOR AFFINES In this section, we discuss the relationship between de Rham and crystalline cohomology (with coefficients) when S is affine. First, we establish some notation that will be used throughout this section. Notation 2.1. Assume S = Spec(A) for a Z=pN -algebra A (and some N > 0). Choose a polynomial algebra P over ^ Zp and a surjection P ! A with kernel J. Let D = DJ (P ) be the p-adically completed divided power envelope of 1 1 ^ 1 e 1 P ! A. We set ΩD := ΩP ⊗P D, so ΩD=p ' ΩD=peD. We also set D(0) = D and let ^ D(n) = DJ(n)(P ⊗Zp · · · ⊗Zp P ) where J(n) = Ker(P ⊗ · · · ⊗ P ! A) and where the tensor product has (n + 1)-factors. For each e ≥ N and any e n ≥ 0, we have a natural object (S; Spec(D(n)=p (n)); δ(n)) of (S=Σ)cris. Using this, for an abelian sheaf F on (S=Σ)cris, we define F(n) := lim F((S; Spec(D(n)=pe(n)); δ(n))): e≥N e e Each (S; Spec(D(n)=p (n)); δ(n)) is simply the (n+1)-fold self-product of (S; Spec(D=p ); δ) in (S=Σ)cris. Letting n vary, we obtain a natural cosimplicial abelian group (or a cochain complex) F(•) := F(0) !F(1) !F(2) ··· ; that we call the Cech-Alexanderˇ complex of F associated to D. 2.2. Some generalities on crystalline cohomology. This subsection collects certain basic tools necessary for work- ing with crystalline cohomomology; these will be used consistently in the sequel. We begin with a brief review of the construction of homotopy-limits in the only context where they appear in this paper. Construction 2.3. Let C be a topos. Fix a sequence T1 ⊂ T2 ⊂ · · · Tn ⊂ · · · of monomorphisms in C. We will construct the functor R limi RΓ(Ti; −); here we follow the convention that G(U) = Γ(U; G) = HomC(U; G) for any pair of objects U; G 2 C. Let AbN denote the category of projective systems of abelian groups indexed by the natural numbers. The functor F 7! limi F(Ti) can be viewed as the composite fΓ(T ;−)g Ab(C) !i i AbN lim!i Ab: Each of these functors is a left exact functor between abelian categories with enough injectives, so we obtain a com- posite of (triangulated) derived functors fRΓ(T ;−)g D+(Ab(C)) !i i D+(AbN) R!limi D+(Ab): We use R limi RΓ(Ti; −) to denote the composite functor. To identify this functor, observe that if we set T = colimi Ti, then F(T ) = limi F(Ti) by adjunction. Moreover, for any injective object I of Ab(C), the projective system i 7! I(Ti) has surjective transition maps I(Ti+1) !I(Ti): the maps Ti ! Ti+1 are injective, and I is an 2 N injective object. Since projective systems in Ab with surjective transition maps are acyclic for the functor limi (by the Mittag-Leffler condition), there is an identification of triangulated functors RΓ(T; −) ' R lim RΓ(Ti; −): i • • • Thus, the value R limi RΓ(Ti; F) is computed by I (T ) = limi I (Ti), where F! I is an injective resolution. An observation that will be useful in the sequel is the following: if each RΓ(Ti; F) is concentrated in degree 0, then j R limi RΓ(Ti; F) coincides with R limi F(Ti), and thus has only two non-zero cohomology groups (as R limi Ai = 0 for j > 1 and any N-indexed projective system fAigi of abelian groups).
Load more

Recommended publications
  • Introduction
    Chapter 1 Introduction In the late 1940s Weil posed the challenge to create a cohomology theory for algebraic varieties X over an arbitrary field k with coefficients in Z. Weil had a deep arithmetic application in mind, namely to prove a conjecture of his concerning the -function attached to a variety over a finite field: Its first part asserts the rationality of this - function, the second claims the existence of a functional equation, and the third predicts the locations of its zeros and poles. Assuming a well-developed cohomology theory with coefficients in Z, possessing in particular a Lefschetz trace formula, he indicated a proof of these conjectures, cf. [48]. As we now know, for fields k of positive characteristic one cannot hope to con- struct meaningful cohomology groups H i .X; Z/. Nevertheless Weil’s quest provided an important stimulus. Starting with Serre and then with the main driving force of Grothendieck, several good substitutes for such a theory have been constructed. Many later developments in arithmetic algebraic geometry hinge crucially on these theories. Among the most important cohomology theories for algebraic varieties over a field k are the following: If k is of characteristic zero, one may pass from k to the field of complex numbers C and consider the singular cohomology of the associated complex analytic space X.C/. Other invariants in the same situation are given by the algebraic de Rham cohomology of X. For general k and a prime ` different from the characteristic of k, there is the cohomology theory of étale `-adic sheaves due to Grothendieck et al.
    [Show full text]
  • Twisted Differential Operators and Q-Crystals Michel Gros, Bernard Le Stum, Adolfo Quirós
    Twisted differential operators and q-crystals Michel Gros, Bernard Le Stum, Adolfo Quirós To cite this version: Michel Gros, Bernard Le Stum, Adolfo Quirós. Twisted differential operators and q-crystals. 2020. hal-02559467 HAL Id: hal-02559467 https://hal.archives-ouvertes.fr/hal-02559467 Preprint submitted on 30 Apr 2020 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Twisted differential operators and q-crystals Michel Gros, Bernard Le Stum & Adolfo Quirós∗ Version of April 29, 2020 Abstract We describe explicitly the q-PD-envelopes considered by Bhatt and Scholze in their recent theory of q-crystalline cohomology and explain the relation with our notion of a divided polynomial twisted algebra. Together with an interpretation of crystals on the q-crystalline site, that we call q-crystals, as modules endowed with some kind of stratification, it allows us to associate a module on the ring of twisted differential operators to any q-crystal. Contents Introduction 2 1 δ-structures 3 2 δ-rings and twisted divided powers 5 3 q-divided powers and twisted divided powers 7 4 Complete q-PD-envelopes 14 5 Complete q-PD-envelope of a diagonal embedding 17 6 Hyper q-stratifications 19 7 q-crystals 22 8 Appendix: 1-crystals vs usual crystals 24 References 25 ∗Supported by grant PGC2018-095392-B-I00 (MCIU/AEI/FEDER, UE).
    [Show full text]
  • A Stacky Approach to Crystals
    A STACKY APPROACH TO CRYSTALS VLADIMIR DRINFELD Abstract. Inspired by a theorem of Bhatt-Morrow-Scholze, we develop a stacky approach to crystals and isocrystals on “Frobenius-smooth” schemes over Fp . This class of schemes goes back to Berthelot-Messing and contains all smooth schemes over perfect fields of char- acteristic p. To treat isocrystals, we prove some descent theorems for sheaves of Banachian modules, which could be interesting in their own right. 1. Introduction Fix a prime p. 1.1. A theorem of Bhatt-Morrow-Scholze. Let X be a smooth scheme over a perfect field k of characteristic p. Let Xperf be the perfection of X, i.e., X := lim(... −→Fr X −→Fr X −→Fr X). perf ←− Let W (Xperf ) be the p-adic formal scheme whose underlying topological space is that of Xperf (or of X) and whose structure sheaf is obtained by applying to OXperf the functor of p-typical Witt vectors. Theorem 1.10 of [BMS] can be reformulated as a canonical isomorphism RΓcris(X, O)= RΓ(W (Xperf)/G , O), where G is a certain flat affine groupoid (in the category of p-adic formal schemes) acting 1 on W (Xperf), and W (Xperf)/G is the quotient stack . If X is affine this gives an explicit complex computing RΓcris(X, O) (the complex is constructed using the nerve of G ). arXiv:1810.11853v1 [math.AG] 28 Oct 2018 One can define G to be the unique groupoid acting on W (Xperf ) such that the morphism G → W (Xperf ) × W (Xperf ) is obtained by taking the divided power envelope of the ideal of the closed subscheme Xperf ×X Xperf ⊂ W (Xperf ) × W (Xperf ).
    [Show full text]
  • Prismatic $ F $-Crystals and Crystalline Galois Representations
    PRISMATIC F -CRYSTALS AND CRYSTALLINE GALOIS REPRESENTATIONS BHARGAV BHATT AND PETER SCHOLZE Abstract. Let K be a complete discretely valued field of mixed characteristic (0,p) with perfect residue field. We prove that the category of prismatic F -crystals on OK is equivalent to the category of lattices in crystalline GK -representations. Contents 1. Introduction 1 2. The absolute prismatic site and the quasi-syntomic site 5 3. Local systems on the generic fibre via the prismatic site 9 4. Prismatic F -crystals 12 5. Prismatic F -crystals over Spf( K ): formulation of the main theorem 15 6. From crystalline Galois representationsO to prismatic F -crystals 19 7. Crystalline Galois representations and Breuil-Kisin modules 28 References 37 1. Introduction Let K be a complete discretely valued field of mixed characteristic (0,p) with ring of integers , perfect OK residue field k, completed algebraic closure C, and absolute Galois group GK . 1.1. The main theorem. A fundamental discovery of Fontaine [Fon82] in the study of finite dimensional Qp-representations of GK is the property of being crystalline; this notion is an (extremely successful) attempt at capturing the property of “having good reduction” for such representations, analogous to the property 1 of “being unramified” in the ℓ-adic case . For instance, to each crystalline GK -representation V , Fontaine has attached an F -isocrystal Dcrys(V ) over k of the same rank as V , thus providing a notion of “special fibre” for such representations, thereby solving Grothendieck’s problem of the mysterious functor. The motivating example here comes from2 algebraic geometry: given a proper smooth algebraic variety X/K arXiv:2106.14735v1 [math.NT] 28 Jun 2021 with good reduction determined by a proper smooth scheme / K extending X/K, each GK -representation i i i X O H (X , Q ) is crystalline and D (H (X , Q )) H ( )Q .
    [Show full text]
  • Barsotti-Tate Groups and Dieudonné Crystals
    BARSOTTI–TATE GROUPS AND DIEUDONNÉ CRYSTALS ALEXANDRE GROTHENDIECK TRANSLATED BY ERIC PETERSON FOREWORDS This is an English translation of Grothendieck’s Groupes de Barsotti–Tate et Cristaux de Dieudonné, published in 1974 by Les Presses de l’Université de Montréal in their Séminaire de Mathématiques Supérieures sequence. While the French original is actually quite legible even to an anglophone, I find it preferable to be able to read this interesting lay-of-the-land document without even the minor distraction of having to mentally pair the cognates “cohomologie” and “cohomology”. I hope that this translation encourages a wider audience to become familiar with its contents, rather than (as was the case for me) skimming the main for the biggest ideas and relegating the rest to a rainy day. I’ve taken some liberties with the prose, but I’ve done my best to preserve the letter of the mathematics. Corrections are warmly welcomed at [email protected]. Eric Peterson These notes do not include certain material treated by Prof. Grothendieck in his 1970 Avertissement summer Montreal course. We hope to include in a later edition those materials not published here. Among the items omitted here, the first is a chapter intended to cover F –crystals, which would have been inserted between the final two. As a compromise, one will instead find a letter from Grothendieck to Barsotti in an appendix at the end of the notes which covers the intended applications of F –crystals. One may also consult the notes of Demazure on p–divisible groups [Dem72], here called Barsotti–Tate groups.
    [Show full text]
  • TANNAKIAN FORMALISM OVER FIELDS with OPERATORS 3 Construction Is Available When M Is Now a Formal Set, and S Is a (Nice) Scheme
    TANNAKIAN FORMALISM OVER FIELDS WITH OPERATORS MOSHE KAMENSKY Abstract. We develop a theory of tensor categories over a field endowed with abstract operators. Our notion of a “field with operators”, coming from work of Moosa and Scanlon, includes the familiar cases of differential and difference fields, Hasse–Schmidt derivations, and their combinations. We develop a cor- responding Tannakian formalism, describing the category of representations of linear groups defined over such fields. The paper extends the previously know (classical) algebraic and differential algebraic Tannakian formalisms. Introduction We study fields with operators (briefly described below, and more thoroughly in §2), and linear groups over such fields. Given such a group G (as defined in §3), our goal is to describe the category RepG of finite dimensional representations of G, in a manner similar to the classical Tannakian formalism. In addition to gen- eralising the usual Tannakian formalism, this paper forms a natural generalisation and reformulation of the theory of differential Tannakian categories (Ovchinnikov [17]), and especially of the definition of differential tensor categories in Kamensky [11, § 4]. We mention that results this kind are expected to have applications to Galois theory of linear equations with various operators. The classical Galois theories of ordinary differential and difference linear equations (as explained in Put and Singer [21] and Put and Singer [20], respectively) may be approached via the classical Tannakian formalism (also in Deligne [7, § 9]). More recently, there are the Galois theory of (linear) partial differential equations (initiated by Cassidy and Singer [5]) to which the differential Tannakian theory mentioned above was applied in Ovchin- nikov [18] (see also Gillet et al.
    [Show full text]
  • A Mini-Course on Crystalline Cohomology
    A mini-course on crystalline cohomology June 15, 2018 Haoyang Guo Abstract This is the lecture notes for the mini-course during June 11-15, 2018 at University of Michi- gan, about the crystalline cohomology. In this mini-course, we will give an overview about crystalline theory. We first give a mild introduction about the motivation and main results of crystalline cohomology, without anything technical. Then we start by looking at algebraic and geometric basics around the crystalline theory. We prove the comparison theorem between crys- talline cohomology and de Rham cohomology, following Bhatt and de Jong [BdJ12]. After that, we turn to the study of the de Rham-Witt complex, a powerful tool in crystalline cohomology. At last, we apply our theory to several questions about rational points in arithmetic geometry.1 Contents 1 Introduction 1 1.1 Motivations . .1 1.3 Foundational results . .4 2 Pd-structure and pd-algebra 5 3 Crystalline geometry 7 4 Crystalline cohomology and de Rham cohomology 10 4.1 Affine case . 10 4.8 Global case . 13 5 De Rham-Witt complex 14 6 Some results of the De Rham-Witt complex 18 1 Introduction 1.1 Motivations Crystalline cohomology was invented by Grothendieck in 1966 , in order to find a "good" p-adic cohomology theory, to fill in the gap at p in the families of `-adic ´etalecohomology, and to refine the notion of the de Rham cohomology in positive characteristic. It was started by Grothendieck's lectures at IHES in 1966 ([Gro68]), where he outlined the program, and was worked out by Berthelot in his thesis ([Ber74]) To motivate the story, we first look at some known cohomology theory at that time.
    [Show full text]
  • Integral Crystalline Cohomology Over Very Ramified Valuation Rings
    JOURNAL OF THE AMERICAN MATHEMATICAL SOCIETY Volume 12, Number 1, January 1999, Pages 117{144 S 0894-0347(99)00273-8 INTEGRAL CRYSTALLINE COHOMOLOGY OVER VERY RAMIFIED VALUATION RINGS GERD FALTINGS 1. Introduction In this paper we intend to show how Fontaine’s comparison-theory ([Fo1], [Fo2]) between crystalline and p-adic ´etale cohomology can be extented to very ramified base rings. So far it had been developed either for Qp-coefficients, or Zp-coefficients if the base is unramified over the Witt-vectors (see [Fa2]). For example it states that for any p-adic discrete valuation-ring V there exists a ring B(V )withthe following property: If X denotes a smooth and proper V -scheme, the p-adic ´etale cohomology of the ¯ generic fiber X V K is related to the crystalline cohomology of X/V0 (V0 defined below) by an isomorphism⊗ ¯ H∗ (X V K,Qp) p B(V )=H∗ (X/V0) V0 B(V ). et ⊗ ⊗Q crys ⊗ One can recover the cohomologies from the above by either taking Frobenius- invariants in filtration degree 0, or Galois invariants. Thus crystalline and ´etale cohomology determine each other. A more general theory might be possible, but so far the attempts to do that have not been entirely successful (see [Fa4]), as the theorem there does not seem to have any applicable consequences. However it turns out that for schemes for which the Hodge cohomology has no torsion, and also for p-divisible groups (in some sense an honorable member of the previous class), one can find meaningful results. A.
    [Show full text]
  • Introduction to Crystalline Cohomology
    Introduction to crystalline cohomology Pedro A. Castillejo 22.10.2015 Abstract This is the second talk of the Research Seminar \Supersingular K3 surfaces are unirational" of the Arithmetic Geometry group (FU Berlin), which is on the Winter Semester 2015-2016. In the seminar we study the result of Liedtke in [Lie15b]: Theorem 1. Supersingular K3 surfaces are unirational. One of the ingredients of the proof is crystalline cohomology, and this talk is devoted to give an introduction to it. In these notes for the talk you can find the following: • We first give a motivation, explaining why crystalline cohomology is like a \p-adic" cohomology, and first we recall the construction and some properties of `-adic cohommology. • After this, we recall the construction and m ain properties of the Witt vector, since crystalline cohomology will have these coefficients. After this, we sheafify this construction as in [Ser58], but the first idea that one may have, which is to take sheaf cohomology with this coefficients, doesn't work. • Hence, we move on to crystalline cohomology, explaining some properties and giving the definition. We also do a couple of examples. • Finally, we look at the de Rham-Witt complex, and state the canoni- cal decomposition of the crystalline cohomology given by the Frobenius action. 1 Motivation What is brown outside, white inside and is very delicate? Crystalline cocohomology. Crystalline cohomology was at first motivated by the search of a cohomology theory analogous to the `-adic cohomology for a scheme over a field of characteristic p, with p 6= `. In fact, under the assumption ` 6= p, `-adic cohomology has a lot of nice properties which become false if we allow ` = p.
    [Show full text]
  • Cris Is for Crystalline Marc-Hubert Nicole 1. Cristalline Topology And
    Cris is for Crystalline Marc-Hubert Nicole 1. Cristalline topology and divided powers Goal: explain the definition of the crystalline site There are two distinct aspects to the definition of the crystalline site: some geomet- ric data and some algebraic data. The algebra, or rather PD-algebra (for Puissances Divis´ees,or divided powers) is a remedy to the non-existence of the usual logarithm in characteristic p. The geometry that arises is nilgeometry, the geometry of infinitesimal thickenings. 1.1. The logarithm in characteristic p. It is a truism in the arithmetic of global fields that function fields are easier to deal with than number fields: the existence of additional structures (e.g. the Frobenius map x 7! xp) gives us more handles, more tools (e.g. the Cartier isomorphism), to prove theorems. One aspect in which positive characteristic is more difficult than characteristic zero is the pathological behaviour of the usual differential calculus (d(xp) = pxp¡1 = 0 in char. p), or, equivalently, the non-existence of a well-behaved logarithm function. Nonetheless, one can cook up useful exponential and logarithm functions in characteristic p, under certain restrictive hypotheses. I know of essentially only one way of circumventing the division by zero problem in characteristic p, and it involves truncating a suitable infinite series giving the logarithm. A primitive version of it comes as follows: Let A be an associative Fp-algebra. Put N := fz 2 Ajzp = 0g; and U := fz 2 Ajzp = 1g: We may define exp and log by: p¡1 p¡1 X zi X (1 ¡ z)i exp(z) = and log(x) = ¡ : i! i i=0 i=1 Then exp : N ¡! U and log : U ¡! N are inverses of each other.
    [Show full text]
  • The Crystalline Site
    The Crystalline Site Matt Tyler The purpose of these notes is to define a crystal on a scheme X. A crystal is an object on the crystalline site of X, so we will first need to define the crystalline site, which in turn requires the notion of divided powers, so it is there that we will begin. The information in these notes comes primarily from Chapter III of Messing [2] and Chap- ters III, V, and VI of Berthelot-Ogus [1]. 1 Divided Powers xn Divided powers give us a way of making sense of expressions of the form n! even in positive characteristic, where it is generally not possible to divide by n!. In this section, we will define divided powers, and discuss some examples and constructions of divided power rings. Definition 1.1. A divided power structure on a ring A and an ideal I/A is a collection of maps γn : I ! A for all n ≥ 0 satisfying the following properties for all x; y 2 I and λ 2 A: (1) γ0(x) = 1, γ1(x) = x, and γn(x) 2 I for all n ≥ 1. P (2) γn(x + y) = i+j=n γi(x)γj(y). n (3) γn(λx) = λ γn(x). (i+j)! 1 (4) γi(x)γj(x) = i! j! γi+j(x). (pq)! 2 (5) γp(γq(x)) = p!(q!)p γpq(x). From (3), we may immediately deduce (6) γn(0) = 0 for all n > 0, and from (1) and (4) and induction on n, we find n (7) n! γn(x) = x .
    [Show full text]
  • Of the American Mathematical Society ISSN 0002-9920
    Notices of the American Mathematical Society ISSN 0002-9920 ABCD springer.com New and Noteworthy from Springer of the American Mathematical Society Ramanujan‘s Lost Notebook Scheduling INCLUDES September 2008 Volume 55, Number 8 Part II Theory, Algorithms, CD-ROM G. E. Andrews , Penn State University, University Park, PA, USA; and Systems B. C. Berndt , University of Illinois at Urbana, IL, USA M. L. Pinedo , New York The primary topics addressed in this second volume on the University, New York, NY, USA lost notebook are q-series, Eisenstein series, and theta This book on scheduling covers both functions. Most of the entries on q-series are located in the theoretical models as well as heart of the original lost notebook, while the entries on scheduling problems in the real world. Eisenstein series are either scattered in the lost notebook or It includes a CD that contains movies with regard to are found in letters that Ramanujan wrote to G.H. Hardy from implementations of scheduling systems as well as slide- nursing homes. shows from the industry. 2008. Approx. 430 p. 3 illus. Hardcover 3rd ed. 2008. With CD-ROM. Hardcover ISBN 978-0-387-77765-8 7 approx. $89.00 ISBN 978-0-387-78934-7 7 $89.95 An Introduction to Semiparallel Homological Algebra 2ND Submanifolds in EDITION J. J. Rotman , University of Illinois at Space Forms Urbana-Champaign, IL, USA Ülo Lumiste, University of Tartu, In this brand new edition the text has been fully updated and Estonia revised throughout and new material on sheaves and abelian Quite simply, this book is the most categories has been added.
    [Show full text]